Select Data with Early Stopping Based on Standard Errors — data_selection • sparrow

Performs greedy data selection with an early stopping criterion when standard errors converge.

Usage

data_selection(
  X,
  max_data_size,
  min_standard_error = NULL,
  log = FALSE,
  period = 1000
)

Arguments

X: Covariate matrix (p x n)
max_data_size: maximum number of data points to select
min_standard_error: Target standard errors for convergence
log: Whether to print progress
period: If log = TRUE, print progress every period

Value

A vector of selected data point indices